I do not need to provide metrics or evidence for this I think anyone who was using composer 1.5 a week ago would know it was a near opus tier model and now it has been reduced to one of the least helpful models I’ve ever dealt with, constantly diverging from task, ignoring instructions, forgetting things even without chat summarization. its disappointing what you guys have done to such a good model. Why would I use this anymore? If it really is such a hurt to run a non-quantized version of composer 1.5 as auto usage than just make it API, I want my damn model back.
Same. Please bring back composer 1 
Same. Please bring back composer 1